Wizard
The Wizard sweeps the asset inventory for problems worth a person's attention, such as the same host recorded twice or a name two sources disagree on, and raises a finding for each. You then fix the inventory (often by merging two assets), mark the finding resolved, or exclude it if it is not a problem.
Open it from Wizard in the sidebar. The screen has two tabs: Findings and Exclusions.
Permissions
| To do this | You need |
|---|---|
| See analyzers, findings and exclusions | Wizard Read |
| Run an analysis, resolve or exclude a finding, remove an exclusion | Wizard Write |
The analyzers
| Analyzer | Offered for | What it looks for |
|---|---|---|
| Duplicate assets | Endpoints, Subnets, Users | One hostname claimed by several assets, usually the same machine discovered twice by different sources |
| Name conflicts | Endpoints, Subnets, Users | Assets whose hostname disagrees with the name a source reports for them, such as a machine renamed after one source cached the old name |
| Invalid names | Users | Assets whose name does not look like a valid hostname |
Duplicate assets
The analyzer groups endpoints by their Names value. Every name held by more
than one endpoint becomes one finding, for example "Hostname web01 belongs to
several (3) assets", with links to the assets (up to five are listed). Open
them, decide which is the real record, and merge the
others into it.
Name conflicts
For each endpoint, the analyzer compares the names in Names with the hostname that each of these sources reports in its own field: Windows Defender, Nessus agents, Elastic agent, CrowdStrike, Carbon Black, Azure managed devices, the active scanner and Active Directory.
Names are compared as bare hostnames, without the domain and ignoring case, so
web01 and WEB01.corp.example match. A source conflicts only when none of the
names it reports is one the asset already goes by. One finding is raised per
endpoint, listing each conflicting source field and its value.
In the current code, Duplicate assets and Name conflicts analyse endpoints whichever asset type is chosen in Analyze, and Invalid names has no analyzer behind it, so a run of it ends as Failed. Confirm the intended behaviour before release and update this page.
Running an analysis
Findings do not appear on their own: an analyzer has to run first.
- Choose Run analysis (top right, or on the empty Findings tab).
- In Run an analysis, pick the asset type under Analyze. The dialog lists the analyzers offered for that type, with what each looks for and the status and time of its last run, or Never run.
- Choose Run beside the analyzer.
The run is queued and processed in the background. Its status reads Queued, Running, Completed or Failed. When it completes, its findings appear on the Findings tab.
Each run also tidies up after itself: a finding from an earlier run of the same analyzer that the new run no longer detects is closed automatically, so fixing the inventory and running again clears the list.
Analyses are carried out by a background worker of your Organization. If runs stay Queued, ask your administrator whether the Wizard worker is enabled for the deployment.
Working with findings
The Findings tab lists the tenant's open findings with the Analyzer, What was found and when it was Detected. In the description, assets are links to their asset page, disputed values are highlighted, and source fields are shown by their display name (the key is in the tooltip).
The search box above the table filters only the page in view, not every page.
Each finding's row menu has two actions. A finding that has already been resolved or excluded cannot be acted on again.
Mark as resolved
Mark as resolved (confirm with Mark resolved) closes the finding and removes it from the list. Resolving is a statement that the problem has been dealt with: if the next run of the analyzer still detects the same problem, it is raised again.
Exclude from future runs
Exclude from future runs is for a finding that is not a problem and never will be, for example a host that is deliberately known by two names.
- Choose Exclude from future runs. The dialog shows the finding.
- Under Why is this not a problem?, write the reason. It is required, up to 1,000 characters.
- Choose Exclude finding.
The exclusion is stored against the finding's fingerprint, so every future run of that analyzer in this tenant skips the same problem. The fingerprint identifies the problem, not the finding: for Duplicate assets it is the shared hostname, for Name conflicts it is the asset.
An exclusion keeps no copy of the finding it came from. The reason you type is the only record of what was waved through and why, so write it for whoever reviews the exclusions later.
Exclusions
The Exclusions tab lists the tenant's exclusions with the Analyzer, the Reason for excluding, the Fingerprint and when it was Excluded. As on the Findings tab, the search box filters only the page in view.
Remove exclusion (in the row menu, confirmed with Remove exclusion) deletes it. The next analysis reports the problem again if it still exists.
See also
- Assets: merging duplicates and editing names
- Asset types: what each list holds
- Integrations: the sources whose names the analyzers compare